by Derek Parfit
Oxford University Press, Volume 1: 540 pp., $35.00; Volume 2: 825 pp., $35.00
1.
Philosophers have long sought to
formulate a theory that explains the purposes of commonsense moral rules
and provides principles enabling us to resolve the frequent moral
dilemmas we encounter. Thomas Hobbes wrote that familiar moral rules are
not relative to one culture or another but are “articles of peace,”
necessary to civilized social life. It is in everyone’s rational
self-interest to obey these rules; the grim alternative is a “state of
war.” Immanuel Kant said that we have an unconditional duty to obey
morality regardless of our desires and self- interests. His second
“categorical imperative” says that we ought never treat others “merely
as means,” but always as “ends in themselves.” To do so, we should
follow a general principle that we believe everyone should follow in
circumstances like our own.
Kant held that his imperative
justifies our commonsense duties to each other and provides a more
fine-grained method of reasoning about what we ought to do when ordinary
moral rules do not adequately address the complexities of life. For
example, under what circumstances is it permissible to break a promise
or deceive someone? To save innocent life or prevent great harm, surely,
but not to benefit ourselves in minor ways; the hard cases lie in
between.
In The Methods of Ethics (1874), among
the greatest works in moral philosophy since Kant, the British
philosopher Henry Sidgwick countered that the rules of commonsense
morality coincide with utilitarianism. The “principle of utility” says
that our actions are right and our laws are just to the degree that they
promote the greatest sum of “utility,” or happiness, in the world.
Happiness, Sidgwick says, is basically pleasurable experiences.
Contemporary
utilitarians often identify happiness (now called “welfare” or
“well-being”) with satisfaction of preferences, or of rationally
informed desires. Until John Rawls’s influential social contract theory,
expounded in A Theory of Justice (1971), utilitarianism remained
the predominant moral theory in Anglo-American philosophy for over two
hundred years. Utilitarianism remains highly influential among
economists, in business and law schools, and in public policy
institutes. In each it is common to hear arguments that a law or
practice is justified because it improves overall well-being.
Utilitarianism
is the most prominent example of a family of positions called
“consequentialism.” These positions hold that actions, laws, or other
conventions are right to the degree that they produce the best
consequences, effectively “maximizing” the good. Many consequentialists
today consider utilitarian general happiness only one of the good
consequences that right conduct ought to promote. Some say that
equalizing the distribution of happiness is also important. Some
consequentialists endorse, as among the “intrinsic goods” that ought to
be promoted, goods such as knowledge, creativity, aesthetic
appreciation, love and friendship, or individual freedom.
Joining
consequentialism and Kantianism is a third major position in
contemporary moral philosophy, Harvard philosopher T.M. Scanlon’s
“contractualism,” which reflects Rawls’s social contract theory of
justice, the main influence on Scanlon. Rawls contends that justice
requires that we act upon principles that would be unanimously agreed to
among free persons equally situated behind an impartial “veil of
ignorance” where they do not know particular facts that would bias their
judgments.
Modifying Rawls’s social contract to apply it
to personal duties, Scanlon’s contractualism says that we owe to each
other a general duty to act on moral rules—such as not harming others
and honoring our promises—that it would be unreasonable for anyone to
reject. Contractualism resembles Kantian views in that it sees the
morality of right and wrong as duties we owe to one another in
recognition of our equal status as persons. In this respect, both stand
together in opposition to consequentialist views, which construe right
and wrong as derived from an impartial duty to promote the best overall
states of affairs in the world, even if in the course of doing so what
contractualists see as moral duties to persons may not be fulfilled.
2.
Consequentialism,
Kantianism, and contractualism are currently the three predominant
positions in moral philosophy, and they are the primary subject of On What Matters, Derek Parfit’s enormous two-volume treatise.1
The book is divided into three main discussions: Part I, “Reasons,”
argues for the objectivity of reasons for acting; Parts II–V are on the
three main moral theories just mentioned; and Part VI, “Normativity,”
defends the truth of moral and other normative judgments. There is also a
helpful introduction by the book’s editor, Samuel Scheffler, and four
critical commentaries by the philosophers Susan Wolf, Barbara Herman,
Allen Wood, and Scanlon.
Parfit’s treatise is driven by
two overarching concerns. First, he hopes to show that moral
philosophy’s three predominant positions converge into a “Triple
Theory.” Parfit’s Triple Theory says, first, that right and wrong are
determined by moral rules that, when generally accepted, “optimize,” or
promote the best overall consequences in the world. Though this sounds
like a form of consequentialism—indeed Parfit calls it “Kantian rule
consequentialism”—he offers both contractualist and Kantian arguments
for it, appealing to the idea that the rules are ones that it would be
unreasonable for anyone to reject, and that we all have reason to
consent to them. Hence the designation “Triple Theory.”
Parfit’s
second main concern is the truth and objectivity of morality and of
reasons and values more generally. He argues that the statements we make
about moral duties and valuable activities are not subjective or
culturally relative, but are objectively true or false. Things are
valuable, independent of whether we desire or value them. It is for
Parfit an objective truth that happiness is good, suffering is bad, and
that “no one could ever deserve to suffer.” If the reasons for moral and
evaluative choices are objective and they justify true statements about
duties and values, then moral and value relativism, subjectivism, and
nihilism must be false. This is the main conclusion of Parts I and VI.
Parfit
uses these two concerns to address the question of “what matters.” He
discusses some of the things that ultimately matter—primarily happiness
and an absence of human and nonhuman suffering. But he is especially
concerned with showing that something must matter, independent of
our subjective and culturally relative beliefs and desires. If there
are no objective reasons or values but only desires and beliefs about
what matters, then there are no truths about morality and what we ought
to do. But then, Parfit contends, nothing can truly matter—regardless of how much we care about it—and we are condemned to nihilism. On What Matters
dryly sets forth countless arguments, but its author is passionate in
his conviction that there must be objective values that give meaning to
our lives in a godless world. It is rare to find an academic
philosophical treatise that sincerely grapples with such cosmic
questions as “whether human history has been worth it,” given all the
suffering that has existed in the world.
3.
In Part I of On What Matters, “Reasons,” Parfit—challenging a fundamental premise of our consumer culture—denies that we have any
reason at all to satisfy our own desires or preferences for their own
sake. He argues the radical position that the mere fact that an action
would promote the satisfaction of some desire is never in itself a
reason for the person to do that action. People can and do desire most
anything. For example, it’s conceivable, Parfit says, that a person
could desire to be in agonizing pain. Surely this desire gives him no
reason to satisfy it by putting his hand in the fire. In order for a
person to have a reason to act as he desires, there must be some feature
of the object of desire that makes it worth desiring. Practicing
the piano in order to play better may be a goal worth desiring.
Producing agonizing pain by burning your hand is not such a goal.
The
position Parfit attacks here is known as “the desire-based theory of
reasons.” It has enormous influence in philosophy, economics, political
science, rational choice theory, and other academic disciplines.
Underlying the desire-based theory is the premise that, in order for us
to have a reason for doing anything, we must be motivated to act;
and this requires a desire that propels our behavior. For our
capacities for reasoning and intellect are, as David Hume said, “inert,”
incapable of moving people to act in the absence of some desire,
whether for wealth or knowledge or power or others’ happiness, for
example.
The account of rationality implicit in economics
and rational choice theory presupposes the desire-based theory of
reasons for acting. In both, by definition it is rational for a person
to maximize his individual utility: that is, to act to satisfy
consistently ordered preferences for what he most wants. Parfit argues
in effect that we have no reason to maximize our utility if we do
so regardless of the objects of our desires. Whether we have reason to
do what we most want depends, instead, upon the value of the objects of
our desires, and the reasons these objects give us for acting. Pleasure,
knowledge, love and friendship, aesthetic appreciation, justice and
equality, and many other ends might be good reasons for acting and hence
worthy of desire. But it’s these objective values themselves, and not
the mere fact that we desire them, that provide us with reasons to
pursue them.2
Economists may say they are insulated from these criticisms, since their task is to explain,
not justify, individual and group behavior. We often talk about “the
reasons” a person had for acting (e.g., Caesar’s reasons for crossing
the Rubicon), referring to the beliefs and desires that cause conduct,
with no moral or evaluative connotations. Parfit and other critics
recognize this causal usage. What they object to is the subtle
transition from a causal to a normative use of “reason for acting,”
which implies what people ought to do. Economists and rational
choice theorists, wittingly or not, make this transition when they say
that a person acts “irrationally” by not maximizing his individual
utility.
Parfit’s position implies that we often do not
have sufficient reasons to act rationally by maximizing our own
utility; and that acting nonrationally (if not irrationally) by refusing
to satisfy certain utility-maximizing preferences (e.g., to steal, or
cheat on taxes, knowing we will not be discovered) can be the best
course of action. If Parfit is right, then ambitious economists perhaps
should abandon their claim that economics is a “science of rational choice” and instead entitle it the “science of consistent, self-interested choice.”
4.
It
is difficult to understand Kant’s second categorical imperative, that
we are never to treat others “merely as means” but always as “ends in
themselves,” when that imperative is isolated from the rest of Kant’s
moral philosophy. Parfit nonetheless tries. He simplifies his task by
focusing exclusively on treating others “merely as means,” disregarding
the idea of treating persons as ends. Clearly Kant cannot mean we should
never rely on others as means to achieve our purposes; for it
would be hard to make it through life without the services of strangers
(grocers, physicians, teachers, garbage collectors, etc.). But
instrumentally relying upon others is different from treating them merely
as means, with no regard for their rights and interests—as if they were
slaves. To rob someone at gunpoint, or transplant her kidneys without
her consent, is to treat her merely as a means. Parfit suggests,
however, that even when we harm others, we do not treat them merely
as means if we deprive them of no more than is necessary to prevent a
greater harm to someone else. Suppose you sacrifice a person’s leg to
save another’s life, while refraining from sacrificing his second leg to
save your computer. Since the victim’s well-being is considered and is
sacrificed only for a greater good, it cannot be said that the victim is
treated “merely as a means.”
This is an unusual
interpretation of Kant. His second categorical imperative is usually
regarded as a statement of the Enlightenment idea that persons,
regardless of their social, religious, or ethnic status, deserve moral
consideration and a fundamental level of respect simply by virtue of
being persons. The traditional interpretation of Kant’s principle is
that to treat others as ends requires that we respect certain moral
rights that prevent their interests from being sacrificed even for the
sake of creating greater overall good. This crucial element is abandoned
in Parfit’s interpretation, since he regards the imperative as a
prohibition against treating others “merely as means,” a prohibition
that can be satisfied simply by taking their interests into
consideration, even if we do not protect them.
The price
of Parfit’s revisionist reading is that we lose any focus on certain
fundamental ideas driving Kant’s moral philosophy. Kant saw everyone as
having an “innate right to freedom,” including the “independence from
being constrained by another’s choice.” This moral right is a direct
consequence of the importance Kant assigns to individual autonomy, and
to respect for others as ends in themselves. Since Parfit gives these
core ideas little attention, individual rights occupy a secondary
position in his “Kantian consequentialism.” Parfit’s primary concern is
promoting individual well-being. The rights and freedoms people have are
regarded as instrumental measures for promoting greater overall
well-being and its appropriate distribution among people. As Parfit’s
commentators argue, this is far from Kant’s own position; it also risks
deflecting attention from the need to protect human rights.
5.
Parfit
argues that Kant’s first categorical imperative is more important. It
says we should act only on a “maxim” or rule that we could will to
become a “universal law” that is accepted and acted upon by everyone.
After examining many possible interpretations of what Kant meant, Parfit
formulates a principle that makes sense to him: “Everyone ought to
follow the principles whose universal acceptance everyone could
rationally will.” This is the “supreme principle of morality” Kant must
have intended, and Parfit endorses it.
Parfit calls this
principle “Kantian Contractualism.” It resembles, he says, T.M.
Scanlon’s contractualism, which he summarizes as “everyone ought to
follow the principles that no one could reasonably reject.” Neither
version of contractualism tells us specifically which principles
of conduct we ought to endorse; instead both provide procedures for
thinking about the substantive principles of conduct we should observe.
Applying these procedures in his own peculiar way, Parfit argues that
both Kantian and Scanlonian contractualism justify “Kantian Rule
Consequentialism”—his own phrase—which says, “everyone ought to follow
the principles whose universal acceptance would make things go best.”
The
assertion that we ought to do what is best may seem a commonplace moral
platitude. But in moral philosophy, this expression has a specific
meaning: that we should do what best promotes good consequences
impartially construed, or that are, as Parfit puts it, “optimific.” This
word refers to the consequentialist principle that right conduct is
purely instrumental, aimed at achieving and maximizing, or “optimizing,”
good consequences. Morality is then a kind of efficiency in promoting
universal good. What is ultimately good is understood as states of
affairs describable without moral concepts about what is right or just.
Thus utilitarians say the happiness of humantity, or even all sentient
beings, is ultimately good, whereas perfectionists such as Aristotle and
Nietzsche say that what is good is achieving excellences of culture and
of character.
Understood this way, Parfit’s claim that
both Kant and contractualism require that we do what is “optimific” or
“impartially best” is highly controversial. Scanlon for example denies
it is true of his contractualism. Most philosophers regard Kantianism
and contractualism as the primary alternatives to
consequentialism. Parfit contends that Kantians, contractualists, and
consequentialists are all “climbing the same mountain on different
sides.” They are all, he says, committed to the Triple Theory: “An act
is wrong just when such acts are disallowed by the principles that are
optimific, uniquely universally willable, and not reasonably
rejectable.” This is, for Parfit, the supreme principle of morality, the
fundamental ground of our moral duties, and the ultimate test of the
morality and justice of all we do.
Like some other
sweeping positions, Parfit’s Triple Theory aims to neutralize its
adversaries. Sidgwick, the main influence on Parfit, also surveyed the
then-leading “methods of ethics” and argued that, suitably pruned, the
more reasonable positions (intuitionism, perfectionism, Kant, and
commonsense morality) all converge upon the doctrine of “universal
hedonism” (Sidgwick’s name for utilitarianism). Subsequent utilitarians
follow Sidgwick’s strategy, but none so inventively as Parfit. Parfit’s
approach to Kant is far more sympathetic and sophisticated. Though, as I
have suggested, he may have misconstrued or rejected fundamental
features of Kant’s view—autonomy, respect for persons as ends, and the
innate right of freedom—Parfit’s consequentialist interpretation of the
categorical imperative will stimulate philosophers for years to come.
6.
To
test our moral intuitions and ultimately to show that it is always
right to act on principles that maximize good consequences, Parfit
relentlessly applies different versions of the so-called “Trolley
Problem” and similar thought experiments that occupy many discussions in
moral philosophy. The simplest version of the Trolley Problem asks
whether it is permissible to throw a switch that redirects a runaway
train in order to save five persons standing on a track, even though you
know that one person on the other track will be killed as a result.
This is Parfit’s “Tunnel” example. Many believe this is morally
permissible, even though it causes another’s death. But what about
pushing a large man off a bridge to trigger a train’s automatic brake in
order to save five people? (Parfit calls this “Bridge.”) Most people
think this would be wrong. But how is it morally different?
One
difference Parfit mentions is that in Bridge, by pushing the man we
directly cause a person’s death as an indispensable means to save five
others—his body is the instrument we use to save them. By contrast, in
Tunnel the death we cause by switching tracks is not instrumental to
saving the five, but an unfortunate albeit “foreseen side-effect”—unlike
Bridge it would be better were the individual who dies in Tunnel not on
the scene at all. Act-consequentialists reject this distinction and
contend that there is no moral difference; for the results are the same
and all that matters in every instance is saving as many lives as we
can.
Parfit rejects act-consequentialism. It implies, in
the example he calls “Transplant,” that a physician should secretly kill
his own patient in order to transplant his organs to save five others.
According to Kantian and Scanlon’s contractualism, no one can rationally
choose or agree to live in a world where physicians can kill their
patients as a means of saving more lives, since this would undermine
trust and the personal nature of the physician–patient relationship.
Parfit accepts this conclusion, but still, in nonmedical emergencies
involving no personal relationships, such as Bridge and Tunnel, he
appears to affirm a more limited version of the consequentialist
principle that everyone is permitted to do whatever saves the most
lives.
Trolley-like examples are subject to endless
variations. Suppose that in Tunnel the person who dies if we hit the
switch is a healthy ten-year-old who has been tied down by five habitual
child-killers on the other track who will be saved by our action. This
added information would make a difference to most people, though perhaps
not to Parfit. He thinks that bad people, even though they may be
restrained or imprisoned to prevent harms to others, nonetheless should
not suffer for their wrongs, since people are not free or “responsible
for their acts in some desert-implying way” (i.e., in a way that implies
that they deserve to suffer or die for their wrongs).
Susan
Wolf comments that there is no single principle underlying our moral
intuitions in Trolley cases. Allen Wood says unrealistic Trolley-like
thought experiments are “worse than useless for moral philosophy,” since
they (at least in Parfit’s use of them) leave out crucial
information—including individuals’ rights, wrongs, and entitlements—and
presuppose that all that matters is the number of lives saved or
goodness and badness of states of affairs. Yet as Scheffler says, Parfit
relies heavily on these examples to defend his book’s most crucial
conclusion: that Kantian contractualism justifies rule
consequentialism—to repeat, “everyone ought to follow the principles
whose universal acceptance would make things go best.” If Parfit’s
commentators are correct, and nothing he says refutes them, it’s
doubtful that Trolley-like thought experiments tell us much about the
general moral principles we should observe.
7.
Most
people believe that we normally ought to keep our promises, that
parents have a duty to care for their children, and that we should not
kill people simply because it’s advantageous. Are these statements
objectively true independent of custom and our subjective attitudes?
Many people, including philosophers, would deny that moral judgments can
be objectively true. They might say that although moral duties are very
important, they are expressions of our emotions, or depict social
conventions, or, as Karl Marx argued, are ideological illusions
obscuring class privileges. Part VI of On What Matters, “Normativity,” is a long defense of the objectivity of moral and other normative concepts and judgments.
Parfit
argues that there are true normative statements about our moral duties
and the values worth pursuing in life. The only alternative he sees is
nihilism. If there were not such truths, “nothing would matter, and
there would not be better or worse ways to live.” Parfit addresses the
metaphysical objection that there can be no moral truths since truth
then would have to depend upon the existence of mysterious moral facts
that have no basis in nature. He argues that normative statements about
values and our duties are true by virtue of the objective reasons
that support them, and that there are objective reasons in the same way
that there are numbers. In order for “2+2=4” to be true, numbers must
exist in some sense, but we needn’t suppose that they are mysterious
entities that exist either in nature or in a non-natural world. The same
is true of moral and evaluative reasons.
Parfit
concludes with an extraordinary chapter on Nietzsche. Nietzsche argued
that morality died with the supposed source of morality’s commands
(God), and that there is no value except what we create by willing it.
Parfit replies that no one (including God) can make something right or
good by commanding or willing it so. “Nor can we make anything matter.”
Things matter because there are objective reasons independent of our
will, choices, and desires, and these reasons justify moral and
evaluative truths. Nietzsche’s attempt to overcome nihilism by grounding
value in our choices about how to live, or in the “will to power” and
commands of a Superman, failed (Parfit says), because arbitrary choices
made without objective reasons cannot make something matter or create
value.
8.
In spite of its 1,400 pages, Parfit’s
book does not adequately address two crucial questions. First, why
should we regard morality and justice as maximization of impersonally
good states of affairs? Second, what is the ultimate good that moral
conduct is to maximize?
Sidgwick provides an answer to
the first question. He says it is “a primary intuition of reason” that
it is “‘right’ and ‘reasonable’” to do what is “ultimately conducive to
universal Good or Happiness.” Sidgwick also says that it is a
“self-evident principle that the good of any one individual is of no
more importance, from the point of view (if I may say so) of the
Universe, than the good of any other.” Parfit approvingly calls this the
“Axiom of Personal Impartiality.”
It takes a refined
philosophical sensibility to conceive of morality as a set of rules that
maximally promote impersonally good states of affairs (e.g., the
maximum happiness of all sentient beings, or maximum achievements of
culture). Many people think of morality differently: as grounded in
relations between persons and the specific duties we owe to one another.
That morality is fundamentally about moral relations between persons,
and not about the relations of persons to states of affairs, is the
intuitive idea behind Kantianism and contractualism. Parfit incorporates
Kant and contractualism as two parts of his Triple Theory. But then he
insists, with little argument beyond his intuitive responses to
Trolley-like examples, that the very rules that are justifiable to all,
and to which all could rationally consent, are the selfsame rules that
maximize good states of affairs.
How does he arrive at
this conclusion? Surely it cannot be an inductive inference drawn from a
few Trolley examples. Instead he says, “If we are asked how we can
recognize such truths, we should appeal, as Sidgwick claims, to our
intuitions.” Like many other non-consequentialists, however, I do not
share Parfit’s abstract intuitions. Take the “Axiom of Personal
Impartiality,” that everyone’s good is equally important. Even assuming
that “no one could ever deserve to suffer,” surely we should not be equally
concerned with promoting the happiness and well-being of sadists and
other evil people (e.g., Hitler, Stalin). Also I do not have a
“normatively indubitable” belief that morality and justice are expedient
ways to optimize good states of affairs. It would clarify matters if
Parfit were more specific about the good that is to be optimized.
Sidgwick said universal happiness is the sole “ultimate good” to be
promoted by all conduct.
Parfit similarly thinks that
happiness is intrinsically good and that all suffering is bad. Unlike
Sidgwick, however, he thinks that the distribution of happiness among
people is important, especially compensation for suffering. He further
says, noncommittedly and without discussion, that “friendship, love,
knowledge, and various achievements may in themselves be good.” He is
ambivalent about whether autonomy, or the freedom to choose and to
determine one’s own life, is intrinsically good. What exactly is the
good state of affairs that should be optimized by our conduct? Parfit
does not answer this crucial question.
9.
Even if
these and other aspects of his account are inadequately defended,
Parfit nonetheless provides the most serious attempt by a
consequentialist to come to grips with Kant and bring him under the
consequentialist umbrella. Parfit carries Sidgwick’s great project—the
assimilation of the major contemporary moral philosophies to some
version of consequentialism—much further than Sidgwick or anyone else.
Parfit’s other major contribution is his defense of the objectivity of
value. Here you don’t have to be a consequentialist to agree with
Parfit; indeed many who are nonconsequentialists would make such a
defense. Parfit challenges the pervasive subjectivism and relativism
that prevails in academic culture and beyond. This is not just a
philosopher’s issue. Every philosophy professor is confronted with
students’ facile moral relativism. That Parfit fiercely challenges such
views is one of the more admirable things about his book. He calls into
question our culture’s peculiarly ambivalent position toward its own
values.
